diff mbox series

drm: rcar-du: Use drmm_encoder_alloc() to manage encoder

Message ID 20210113170253.443820-1-kieran.bingham+renesas@ideasonboard.com (mailing list archive)
State New, archived
Headers show
Series drm: rcar-du: Use drmm_encoder_alloc() to manage encoder | expand

Commit Message

Kieran Bingham Jan. 13, 2021, 5:02 p.m. UTC
The encoder allocation was converted to a DRM managed resource at the
same time as the addition of a new helper drmm_encoder_alloc() which
simplifies the same process.

Convert the custom drm managed resource allocation of the encoder
with the helper to simplify the implementation, and prevent hitting a
WARN_ON() due to the handling the drm_encoder_init() call directly
without registering a .destroy() function op.

Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
---
 drivers/gpu/drm/rcar-du/rcar_du_encoder.c | 31 +++++------------------
 1 file changed, 6 insertions(+), 25 deletions(-)

Comments

Laurent Pinchart Jan. 13, 2021, 10:45 p.m. UTC | #1
Hi Kieran,

Thank you for the patch.

On Wed, Jan 13, 2021 at 05:02:53PM +0000, Kieran Bingham wrote:
> The encoder allocation was converted to a DRM managed resource at the
> same time as the addition of a new helper drmm_encoder_alloc() which
> simplifies the same process.
> 
> Convert the custom drm managed resource allocation of the encoder
> with the helper to simplify the implementation, and prevent hitting a
> WARN_ON() due to the handling the drm_encoder_init() call directly
> without registering a .destroy() function op.
> 
> Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")

We could equally point to the patch that has added drmm_encoder_alloc(),
but I'm fine taking the blame :-)

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> ---
>  drivers/gpu/drm/rcar-du/rcar_du_encoder.c | 31 +++++------------------
>  1 file changed, 6 insertions(+), 25 deletions(-)
> 
> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
> index ba8c6038cd63..ca3761772211 100644
> --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
> +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
> @@ -48,21 +48,12 @@ static unsigned int rcar_du_encoder_count_ports(struct device_node *node)
>  static const struct drm_encoder_funcs rcar_du_encoder_funcs = {
>  };
>  
> -static void rcar_du_encoder_release(struct drm_device *dev, void *res)
> -{
> -	struct rcar_du_encoder *renc = res;
> -
> -	drm_encoder_cleanup(&renc->base);
> -	kfree(renc);
> -}
> -
>  int rcar_du_encoder_init(struct rcar_du_device *rcdu,
>  			 enum rcar_du_output output,
>  			 struct device_node *enc_node)
>  {
>  	struct rcar_du_encoder *renc;
>  	struct drm_bridge *bridge;
> -	int ret;
>  
>  	/*
>  	 * Locate the DRM bridge from the DT node. For the DPAD outputs, if the
> @@ -101,26 +92,16 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
>  			return -ENOLINK;
>  	}
>  
> -	renc = kzalloc(sizeof(*renc), GFP_KERNEL);
> -	if (renc == NULL)
> -		return -ENOMEM;
> -
> -	renc->output = output;
> -
>  	dev_dbg(rcdu->dev, "initializing encoder %pOF for output %u\n",
>  		enc_node, output);
>  
> -	ret = drm_encoder_init(&rcdu->ddev, &renc->base, &rcar_du_encoder_funcs,
> -			       DRM_MODE_ENCODER_NONE, NULL);
> -	if (ret < 0) {
> -		kfree(renc);
> -		return ret;
> -	}
> +	renc = drmm_encoder_alloc(&rcdu->ddev, struct rcar_du_encoder, base,
> +				  &rcar_du_encoder_funcs, DRM_MODE_ENCODER_NONE,
> +				  NULL);
> +	if (!renc)
> +		return -ENOMEM;
>  
> -	ret = drmm_add_action_or_reset(&rcdu->ddev, rcar_du_encoder_release,
> -				       renc);
> -	if (ret)
> -		return ret;
> +	renc->output = output;
>  
>  	/*
>  	 * Attach the bridge to the encoder. The bridge will create the
Geert Uytterhoeven Jan. 14, 2021, 10 a.m. UTC | #2
On Wed, Jan 13, 2021 at 6:02 PM Kieran Bingham
<kieran.bingham+renesas@ideasonboard.com> wrote:
> The encoder allocation was converted to a DRM managed resource at the
> same time as the addition of a new helper drmm_encoder_alloc() which
> simplifies the same process.
>
> Convert the custom drm managed resource allocation of the encoder
> with the helper to simplify the implementation, and prevent hitting a
> WARN_ON() due to the handling the drm_encoder_init() call directly
> without registering a .destroy() function op.
>
> Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>

Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
As in "the WARNING from drm_encoder_init() is gone".

Gr{oetje,eeting}s,

                        Geert
Kieran Bingham Jan. 15, 2021, 1:27 p.m. UTC | #3
Hi Laurent,

On 13/01/2021 22:45, Laurent Pinchart wrote:
> Hi Kieran,
> 
> Thank you for the patch.
> 
> On Wed, Jan 13, 2021 at 05:02:53PM +0000, Kieran Bingham wrote:
>> The encoder allocation was converted to a DRM managed resource at the
>> same time as the addition of a new helper drmm_encoder_alloc() which
>> simplifies the same process.
>>
>> Convert the custom drm managed resource allocation of the encoder
>> with the helper to simplify the implementation, and prevent hitting a
>> WARN_ON() due to the handling the drm_encoder_init() call directly
>> without registering a .destroy() function op.
>>
>> Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
> 
> We could equally point to the patch that has added drmm_encoder_alloc(),
> but I'm fine taking the blame :-)

Perhaps, we could point there indeed, I'm surprised that patch/series
didn't seem to add any users of drmm_encoder_alloc() as far as I can see.

I don't think this is a "blame" though. Just a reference to the most
relevant change.

> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
> 

Thanks.

Kieran


>> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
>> ---
>>  drivers/gpu/drm/rcar-du/rcar_du_encoder.c | 31 +++++------------------
>>  1 file changed, 6 insertions(+), 25 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
>> index ba8c6038cd63..ca3761772211 100644
>> --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
>> @@ -48,21 +48,12 @@ static unsigned int rcar_du_encoder_count_ports(struct device_node *node)
>>  static const struct drm_encoder_funcs rcar_du_encoder_funcs = {
>>  };
>>  
>> -static void rcar_du_encoder_release(struct drm_device *dev, void *res)
>> -{
>> -	struct rcar_du_encoder *renc = res;
>> -
>> -	drm_encoder_cleanup(&renc->base);
>> -	kfree(renc);
>> -}
>> -
>>  int rcar_du_encoder_init(struct rcar_du_device *rcdu,
>>  			 enum rcar_du_output output,
>>  			 struct device_node *enc_node)
>>  {
>>  	struct rcar_du_encoder *renc;
>>  	struct drm_bridge *bridge;
>> -	int ret;
>>  
>>  	/*
>>  	 * Locate the DRM bridge from the DT node. For the DPAD outputs, if the
>> @@ -101,26 +92,16 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu,
>>  			return -ENOLINK;
>>  	}
>>  
>> -	renc = kzalloc(sizeof(*renc), GFP_KERNEL);
>> -	if (renc == NULL)
>> -		return -ENOMEM;
>> -
>> -	renc->output = output;
>> -
>>  	dev_dbg(rcdu->dev, "initializing encoder %pOF for output %u\n",
>>  		enc_node, output);
>>  
>> -	ret = drm_encoder_init(&rcdu->ddev, &renc->base, &rcar_du_encoder_funcs,
>> -			       DRM_MODE_ENCODER_NONE, NULL);
>> -	if (ret < 0) {
>> -		kfree(renc);
>> -		return ret;
>> -	}
>> +	renc = drmm_encoder_alloc(&rcdu->ddev, struct rcar_du_encoder, base,
>> +				  &rcar_du_encoder_funcs, DRM_MODE_ENCODER_NONE,
>> +				  NULL);
>> +	if (!renc)
>> +		return -ENOMEM;
>>  
>> -	ret = drmm_add_action_or_reset(&rcdu->ddev, rcar_du_encoder_release,
>> -				       renc);
>> -	if (ret)
>> -		return ret;
>> +	renc->output = output;
>>  
>>  	/*
>>  	 * Attach the bridge to the encoder. The bridge will create the
>
Geert Uytterhoeven March 1, 2021, 10:30 a.m. UTC | #4
On Thu, Jan 14, 2021 at 11:00 AM Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> On Wed, Jan 13, 2021 at 6:02 PM Kieran Bingham
> <kieran.bingham+renesas@ideasonboard.com> wrote:
> > The encoder allocation was converted to a DRM managed resource at the
> > same time as the addition of a new helper drmm_encoder_alloc() which
> > simplifies the same process.
> >
> > Convert the custom drm managed resource allocation of the encoder
> > with the helper to simplify the implementation, and prevent hitting a
> > WARN_ON() due to the handling the drm_encoder_init() call directly
> > without registering a .destroy() function op.
> >
> > Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
> > Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
> > Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
>
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> As in "the WARNING from drm_encoder_init() is gone".

JFYI, the WARNING is now in v5.12-rc1, the fix isn't.

Gr{oetje,eeting}s,

                        Geert
Kieran Bingham March 1, 2021, 11:52 a.m. UTC | #5
Laurent,



On 01/03/2021 10:30, Geert Uytterhoeven wrote:
> On Thu, Jan 14, 2021 at 11:00 AM Geert Uytterhoeven
> <geert@linux-m68k.org> wrote:
>> On Wed, Jan 13, 2021 at 6:02 PM Kieran Bingham
>> <kieran.bingham+renesas@ideasonboard.com> wrote:
>>> The encoder allocation was converted to a DRM managed resource at the
>>> same time as the addition of a new helper drmm_encoder_alloc() which
>>> simplifies the same process.
>>>
>>> Convert the custom drm managed resource allocation of the encoder
>>> with the helper to simplify the implementation, and prevent hitting a
>>> WARN_ON() due to the handling the drm_encoder_init() call directly
>>> without registering a .destroy() function op.
>>>
>>> Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
>>> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
>>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
>>
>> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
>> As in "the WARNING from drm_encoder_init() is gone".
> 
> JFYI, the WARNING is now in v5.12-rc1, the fix isn't.

Does this patch go through you to get into the DRM tree? Or do I need to
re-send it to someone else?

> Gr{oetje,eeting}s,
> 
>                         Geert
>
Laurent Pinchart March 2, 2021, 9:23 p.m. UTC | #6
Hi Kieran,

On Mon, Mar 01, 2021 at 11:52:26AM +0000, Kieran Bingham wrote:
> On 01/03/2021 10:30, Geert Uytterhoeven wrote:
> > On Thu, Jan 14, 2021 at 11:00 AM Geert Uytterhoeven wrote:
> >> On Wed, Jan 13, 2021 at 6:02 PM Kieran Bingham wrote:
> >>> The encoder allocation was converted to a DRM managed resource at the
> >>> same time as the addition of a new helper drmm_encoder_alloc() which
> >>> simplifies the same process.
> >>>
> >>> Convert the custom drm managed resource allocation of the encoder
> >>> with the helper to simplify the implementation, and prevent hitting a
> >>> WARN_ON() due to the handling the drm_encoder_init() call directly
> >>> without registering a .destroy() function op.
> >>>
> >>> Fixes: f5f16725edbc ("drm: rcar-du: Use DRM-managed allocation for encoders")
> >>> Reported-by: Geert Uytterhoeven <geert+renesas@glider.be>
> >>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> >>
> >> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> >> As in "the WARNING from drm_encoder_init() is gone".
> > 
> > JFYI, the WARNING is now in v5.12-rc1, the fix isn't.
> 
> Does this patch go through you to get into the DRM tree? Or do I need to
> re-send it to someone else?

I can send a pull request with this for the -fixes branch, but as
there's a single patch, maybe David or Daniel could pick it up ?
diff mbox series

Patch

diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
index ba8c6038cd63..ca3761772211 100644
--- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
+++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c
@@ -48,21 +48,12 @@  static unsigned int rcar_du_encoder_count_ports(struct device_node *node)
 static const struct drm_encoder_funcs rcar_du_encoder_funcs = {
 };
 
-static void rcar_du_encoder_release(struct drm_device *dev, void *res)
-{
-	struct rcar_du_encoder *renc = res;
-
-	drm_encoder_cleanup(&renc->base);
-	kfree(renc);
-}
-
 int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			 enum rcar_du_output output,
 			 struct device_node *enc_node)
 {
 	struct rcar_du_encoder *renc;
 	struct drm_bridge *bridge;
-	int ret;
 
 	/*
 	 * Locate the DRM bridge from the DT node. For the DPAD outputs, if the
@@ -101,26 +92,16 @@  int rcar_du_encoder_init(struct rcar_du_device *rcdu,
 			return -ENOLINK;
 	}
 
-	renc = kzalloc(sizeof(*renc), GFP_KERNEL);
-	if (renc == NULL)
-		return -ENOMEM;
-
-	renc->output = output;
-
 	dev_dbg(rcdu->dev, "initializing encoder %pOF for output %u\n",
 		enc_node, output);
 
-	ret = drm_encoder_init(&rcdu->ddev, &renc->base, &rcar_du_encoder_funcs,
-			       DRM_MODE_ENCODER_NONE, NULL);
-	if (ret < 0) {
-		kfree(renc);
-		return ret;
-	}
+	renc = drmm_encoder_alloc(&rcdu->ddev, struct rcar_du_encoder, base,
+				  &rcar_du_encoder_funcs, DRM_MODE_ENCODER_NONE,
+				  NULL);
+	if (!renc)
+		return -ENOMEM;
 
-	ret = drmm_add_action_or_reset(&rcdu->ddev, rcar_du_encoder_release,
-				       renc);
-	if (ret)
-		return ret;
+	renc->output = output;
 
 	/*
 	 * Attach the bridge to the encoder. The bridge will create the